A key United States Senate committee has moved proposed military AI guardrails one step closer to becoming law, approving legislation that would regulate how the Pentagon develops, tests, and deploys artificial intelligence and autonomous weapons systems.
The Senate Armed Services Committee endorsed the new framework in its draft of the fiscal year 2027 National Defense Authorization Act, finishing markup of the bill on June 10. The provisions emphasize human judgment and ultimate human responsibility while largely endorsing the Pentagon's embrace of autonomous technologies, according to the Arms Control Association, which reported the details in July 2026. For more context on this story, see our ongoing AI industry coverage.
The proposed legislation would require the Department of Defense to ensure that personnel exercise appropriate levels of human judgment over the use of AI-enabled weapons, and that relevant systems are designed so commanders and operators retain ultimate human responsibility over the use of force.
Codifying Human Oversight
The framework builds on a principle already established in Department of Defense Directive 3000.09, which governs autonomy in weapons systems but has lacked a clear definition of what constitutes meaningful human judgment. The new language seeks to elaborate on that principle in concrete terms.
Under the proposed provisions, autonomous systems would be required to allow for supervision by human operators, include methods for intervention or termination, incorporate fail-safe mechanisms to enable manual control, and provide adequate monitoring data to controllers. The systems would also have to maintain records of target selection data and logic, and operate in compliance with international law, the law of armed conflict, applicable treaties, and Department of Defense policy.
The legislation would codify the existing review process mandated by the 2023 directive while adding verification, validation, testing, and evaluation criteria. The Pentagon's office of the Director, Operational Test and Evaluation, which has had its proposed budget cut by two-thirds in the current budget request, would be charged with ensuring that covered systems function as anticipated against adaptive adversaries.
A Nuclear Red Line
One of the most significant provisions would categorically prohibit the use of AI for any decision to initiate the launch or detonation of a nuclear weapon. That restriction was sought on a bipartisan basis in 2023 by Senator Ed Markey, Representative Ted Lieu, Representative Don Beyer, and Representative Ken Buck.
However, the Senate committee text does not include a ban on using AI for the targeting of nuclear weapons, a step that Senator Kirsten Gillibrand of New York had proposed in a stand-alone bill introduced on June 2. The omission highlights the boundaries of what lawmakers were willing to restrict, even as they moved to constrain the most extreme autonomous capabilities.
The committee would also establish an incident repository for tracking data on system failures, unintended behavior, and near-miss events, a measure aimed at improving accountability as the military integrates more AI into its operations.
A Bipartisan Foundation
The framework largely tracks a bill introduced on June 8 by Democratic leadership, represented by ranking member Jack Reed of Rhode Island and Senator Chris Coons of Delaware, who serves as the top Democrat on the Senate's defense appropriations subcommittee. The text, which was accepted by the Republican majority into the base draft of the policy bill, also incorporates substantial portions of a proposal by Senator Mark Kelly of Arizona and Representative Suhas Subramanyam of Virginia.
Several other Democratic senators had introduced competing proposals in the weeks before the markup, including measures by Senator Adam Schiff, Senator Elissa Slotkin, and Senator Kelly. The Senate committee language based on the Coons-Reed bill is less specific in its protections against domestic surveillance than some of those alternatives.
Elsewhere in the draft bill, the committee recommends creating a robotics and autonomous systems command responsible for force generation, joint training, interoperability, doctrine development, and operational employment through other combatant commands.
Weighing the Impact on Warfighters
Reflecting bipartisan concerns about the effects of rapid AI adoption, a separate section of the draft incorporates a bill proposed by Senator Kelly and Senator Tom Cotton, an Arkansas Republican, that would instruct the Pentagon to assess how AI use affects the maintenance and retention of essential warfighter skills.
The advances come as the United States races to integrate artificial intelligence into its military faster than adversaries, while lawmakers attempt to establish guardrails that prevent catastrophic failures without slowing deployment. The bill still must pass the full Senate and House before reaching the president's desk, and the provisions could change during negotiations between the chambers.
Still, the committee's approval marks one of the most concrete steps yet by Congress to place binding legal limits on how the military uses autonomous systems, setting parameters that could shape the future of warfare for years to come.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
